In my admittedly biased opinion, Service Oriented Architecture (the politically correct buzzword for systems that are built using the WS-* stack) has a reasonable chance to succeed where DCE, CORBA, Jini, a nearly infinite variety of in-house frameworks and even the Internet have failed to deliver on the promises of distributed computing.
My use of “failure” in this case should not be construed to convey any stigma, because distributed computing is inherently difficult, and the [booming voice]_Promise of Distributed Computing_[/booming voice] has been outlandishly overstated from Day 0.
There are approximately as many distributed computing frameworks as there are reasons for their failure. Many have been underspecified, and many more have been underimplemented.
It is my contention that SOA is the first distributed computing framework to attempt to comprehensively address all three problem areas of distributed systems:
- Application Semantics (data formats, protocols, business logic, …)
- Identity (authentication, authorization, …)
- Networking (connectivity, scalability, fault tolerance, physical security, …)
In addition to this breadth of conceptual scope, there are many cultural ingredients baked into SOA that will help it achieve wider adoption than previous distributed computing frameworks.
I won’t even try to back up my assertions just yet, but hopefully I’ll be writing more on the subject soon.